The Ronda peridotites, the small Ceuta body, the Beni Bousera massif, and the deep-seated peridotite bodies inferred from the Bouguer gravity anomalies would have formed part of a large allochthonous mantle slab imbricated in the units of the Alpujarride complex, as supported by the petrological and geochemical similarities of the peridotite massifs (Gervilla et al., 1988, Targuisti, 1994). The abundance of major and minor peridotite bodies in the Gibraltar arc is a consequence of the large-scale fragmentation of the former slab. The individualization and dispersion of these peridotite bodies is congruent with the Miocene extensional directions that are well established in the Betic and Rif chain (García-Dueñas et al., 1992; Martínez-Martínez and Azañón, 1997).

The presence of HP above and below the peridotite slab and the relationship of the PT conditions with the relative position in the Apujarride sequence, not with the relative position with respect to the peridotite slab, indicate that the intracrustal emplacement was later than the HP metamorphism. Thus, HP metamorphism has no direct relationship with the peridotite emplacement as suggested Tubía et al. (1997), but probably with a very high-pressure record in the peridotites (Davies et al. 1993). This fact disagrees with tectonic models that suppose a single and continuous extensional episode after a collisional event until the present-day Alborán basin (see, for example, Platt and Vissers, 1989, Van der Wal and Vissers, 1993;Vissers et al., 1995; Zeck, 1997). These models generally propose radial dispersion and/or emplacement of the peridotite bodies from a central source to explain the present-day distribution around the Gibraltar arc, but they do not take into account the N-S predominant direction of the shear zones that fragment the former slab. In accordance with these models, for example, the Ceuta peridotite slice would be related with top-to-west detachments, not with the N-S (N 190°E) lineated ductile shear zone that encloses the thin relics of peridotites (Kornprobst, 1962; Sánchez-Gómez et al, 1995).

Other models, focused on the peridotite emplacement, consider a tarnspressive main fault (Tubía and Cuevas, 1986), or oblique subduction (Tubía, 1994; Tubía et al., 1997). These models, which are based mainly on detailed observations in the Alpujata massif (Fig. 3), better reflect the structural relationship of an early orogenic configuration, but contemplate different emplacement faults for the Betic and Rif peridotites in a geographic position not far from their present location. Late extension and, more crucially, lower Miocene vertical axis rotations (Saddiqi et al., 1995; Feinberg et al., 1996) invalidate most of their regional implications, such as the meaning of the thrust lineations.

A more reliable general setting has been proposed by Lonergan and White (1997), who suggest a former collision at 35 Ma around the present Balear Islands, far from the current position of the Alboran domain, and then a subduction roll-back process up to the present. The space and timing constraints of this model agree with structural reconstructions (García-Dueñas et al., 1993), although it does not explain in detail the tectonometamorphic history of the Alboran domain.